Unlocking Value through Flow Management

Organizations and teams exist with the primary purpose of transforming creativity, skills, resources, and energy into value for stakeholders. We refer to the process of transitioning from potential value to actual value as “Flow”. Essentially, Flow represents the movement of work items through your product/service development and deployment process, which should be maintained as stable and predictable as possible while continuously improving it. To achieve this, measurement is crucial, and actions must be taken based on the results obtained from these measurements. Therefore, if we aim to manage Flow to support strategic objectives, it is essential to:

  • switch from measuring people to measuring the work, and its flow through the value stream;
  • measure flow at the appropriate levels (teams, portfolio, product family, program, etc.) and take corresponding actions based on the measurement outcomes.

Metrics and forecasting for managing Flow

A pattern sequence for flow management and forecasting

The approach is not limited to any specific method or framework. It can be applied to Lean, Scrum, Kanban, XP, SAFe, LeSS, or any other framework with the necessary contextual adaptations.

To implement this approach, you will need a product backlog management system that supports your operations and value stream. This system should include a mechanism for constraining work in process and tracking the progress of product backlog items.

Once you have established the necessary product backlog management system, you can develop your own flow measurement framework. This framework should be based on early detection signals and promptly addressing any blockages that arise.

Flow Metrics

The fundamental flow metrics consist of Work In Process (WIP), Throughput, Cycle Time, and Work Item Age, which can be seen as a variant of Cycle Time. These metrics are interconnected by a mathematical law, as demonstrated by mathematician John C. Little.

The law, in the version adopted in lean-agile contexts states that WIP = CT * THP meaning that average Work in Process equals average Cycle Time multiplied average Throughput. Behind the apparent simplicity there is quite a complex math, you can dig deeper starting from these MIT paper: Little’s law 50 years later.

But in order to apply Little’s Law to manage flow, you simply need to grasp its fundamental assumptions regarding system stability and the implications it has on your work processes. Let me recap them here:

  • Adding WIP will increase Cycle Times (on average)
  • Do not initiate more work than you can complete
  • Finish all the work that you start
  • Check Items age and the other metrics, and act upon it
  • Engage in continuous forecasting and take action based on the outcomes

While these metrics serve as foundational measures, they are by no means the only ones you can incorporate. They are akin to a speedometer and odometer in a car, and you are encouraged to supplement and enhance them with other relevant measurements.

Continuous forecasting

A Montecarlo simulation of 10.000 throughput scenarios

I use Monte Carlo simulations to forecast progress toward goals within my teams and programs. While that may sound impressive, it may not necessarily provide value to your stakeholders. To make it truly valuable, you will need to:

  • Apply your forecasting at the right level, where you want to see improvements
  • Move from one-time estimates to ongoing forecasting. Continuously update your forecasts as your work progresses along the value stream
  • Move from relying on deterministic forecasts to utilizing probabilistic ones, taking into account various scenarios and their respective probabilities of realization.
  • Employ forecast outcomes as an extra indicator for early detection and proactively respond to them

Managing blockades

Measurements and forecasting hold no value unless you learn and take action based on them. if you implemented the metrics and the forecasting described here, you now have a data-driven early detection system that serves as a foundation for steering and enhancing your daily operations. Use it!

Examples of responses to early detection signals will vary depending on your specific context. They can range from pairing in a software team to assembling managers from different units to visualize progress and obstacles on a board, to redefining your operational strategy, and so forth.

Caveat

One caveat, however, is that the concept of measurement can be overwhelming for some, particularly in more traditional environments. Therefore, my suggestion is to start from your current position. Begin by measuring what falls within your span of influence and respond accordingly to those measurements. Share your findings at any level and demonstrate the tangible and potential benefits of measurement throughout the organization. Make it clear that you are tracking the workflow, not the people. With time, the value will become apparent and rewarding

Conclusions

Since embarking on our data-driven journey, we have observed the following advantages: reduced cycle times and increased throughput rates, enhanced transparency, improved focus, and simplified prioritization. Furthermore, the adoption of a data-driven mindset has fostered meaningful discussions at all levels regarding the potential of flow management to align with strategic goals, drive competitive advantage, and foster value creation.